Re: [Algorithms] General purpose task parallel threading approach

[Nicholas \Indy\ R. <ar...@gm...>]

On Thu, Apr 9, 2009 at 4:25 PM, Sebastian Sylvan
<seb...@gm...> wrote:
> And of course, general code doesn't have a static call graph, so it's even
> impossible to do this analysis unless your code happens to be simple enough
> (see: the halting problem).

I've found the barrier for "simple enough" to actually be rather high,
especially when writing special case analysis (dehydra has been used
in mozilla to build simple call graphs
https://developer.mozilla.org/en/Dehydra). But often it turns out that
building a call graph for optimization during a profile like stage is
very effective, It's easy to get real call data, and you don't have to
worry about the halting problem.

Also, a description of the halting problem with some humor value:
http://www.lel.ed.ac.uk/~gpullum/loopsnoop.html

Nicholas "Indy" Ray